[FieldTrip] Question about minimum norm estimate pipeline

Sat Jun 1 20:44:04 CEST 2013

Hi Arjen,

Thanks for your message. I did align the mri to Talairach; as you can see
from http://i.imgur.com/26nyHYZ.png, the volume conduction model and
sourcespace are both expressed in the same coordinate system (i.e.,
everything's pointing in the same direction) but they're just not sitting
on top of one another. If anyone has any ideas on where that problem was
introduced (or how to re-align them now), I would greatly appreciate it.
Below I have some more details about how I processed that data, if it helps.

I'm trying to go through the data one step at a time and track where the
problem might have happened. When I compare the sourcespace before having
applied any transformation (i.e., the headshape from
<subject>-oct-6-src.fif) to the original mri (orig-nomask.mgz), they look
ok (I don't know how to plot them together, but see
http://i.imgur.com/LGW7YnJ.png and http://i.imgur.com/JUoTxc9.png -- things
at least look like they're on more or less the same plane). Then I
re-register the mri to CTF (
http://fieldtrip.fcdonders.nl/tutorial/minimumnormestimate#source_modelco-registration_of_the_source_space_to_the_sensor-based_head_coordinate_system);
after that, in mri_nom_ctf, the axes are all going in the right direction
but the whole head is tilted forward and the origin of the axes is no
longer at the anterior commisure (see http://i.imgur.com/CTZNOTk.png for
the realigned MRI).  Applying the transformation matrix T to the
sourcespace also seems to tilt it like that (http://i.imgur.com/bcNIpf3.png),
although as can be seen from the first image in this message it doesn't
quite line up with the volume conduction model in the end. As for the
volume conduction model, here is what it looks like at first (
http://i.imgur.com/ZX3m38b.png) and here is what it looks like after
applying the transformation matrix (http://i.imgur.com/vXa3Cnc.png).
Obviously the transformation matrix is doing something, but it's not
getting the sourcespace and volume conduction model lined up; since it's
the same transformation matrix, all I can guess is that there was some
pre-existing difference between the source mesh (the .fif file) and the
anatomical mri (orig-nomask.mgz), but I'm not sure when that came in.

Another minor issue: when I first compared the volume conduction model and
the sourcespace, they were expressed in different units even though I
followed the code in the tutorial. See http://i.imgur.com/orwgcTJ.png: the
sourcespace looks 10x smaller than the volume conduction model, which I
assume is because it is expressed in cm whereas the volume conduction model
is expressed in mm. To get the figure linked at the very beginning of this
message, I had to convert the units of the volume conduction model to cm,
even though that's not in the tutorial.

I notice that the tutorial on the wiki hasn't been edited since October
2012 (other than a few edits I made this month which were just correcting
typos in the prose). Is it possible that what's on the wiki is out of date?
(Also cc'ing Lilla on this.)

Thanks,
Steve

> Message: 1
> Date: Fri, 31 May 2013 08:11:23 +0200 (CEST)
> From: "Stolk, A." <a.stolk at fcdonders.ru.nl>
> To: FieldTrip discussion list <fieldtrip at science.ru.nl>
> Subject: Re: [FieldTrip] Question about minimum norm estimate pipeline
> Message-ID:
>         <
1914237354.1292588.1369980683984.JavaMail.root at sculptor.zimbra.ru.nl>
> Content-Type: text/plain; charset="utf-8"
>
> Hi Steve, A quick guess; did you correctly align your resliced mri to
Talairach space by indicating the commissures (
http://imaging.mrc-cbu.cam.ac.uk/imaging/FindingCommissures ) and, if I'm
correct, a point in the same place, e.g. between the hemispheres? This
should update the transformation matrix. Best regards, Arjen -----
Oorspronkelijk bericht -----
> > Van: "Stephen Politzer-Ahles" <politzerahless at gmail.com>
> > Aan: fieldtrip at science.ru.nl
> > Verzonden: Vrijdag 31 mei 2013 05:53:45
> > Onderwerp: [FieldTrip] Question about minimum norm estimate pipeline
> > Hello all,
> > I have not yet gotten a response to my question below, but in the
> > meantime I have another question about the minimum norm estimate
> > workflow--specifically, about the coordinate system for the
> > skull-stripped anatomy in the step described at
> >
http://fieldtrip.fcdonders.nl/tutorial/minimumnormestimate#preprocessing_of_the_anatomical_mrisave_to_disk
> > . I'm confused by the following bit of code:
> > % ensure that the skull-stripped anatomy is expressed in the same
> > coordinate system as the anatomy
> > seg.transform = mri_tal.transform;
> > In my data, mri_tal.coordsys is 'spm' (I presume this is the result of
> > re-aligning to Talairach in the previous step?) whereas seg.coordsys
> > is 'ctf' (as a result of re-aligning to CTF several steps earlier).
> > (But mri_tal also has a field mri_tal.transformorig, which seg does
> > not have.) So should I really be using the same transform for both, as
> > shown in the tutorial?
> > Apologies if this question is pretty basic; I'm just trying to
> > pinpoint where the mis-alignment described in my message below
> > occurred, so I want to make sure I understand each step of the
> > workflow correctly
> > Best,
> > Steve
> > > Message: 1
> > > Date: Sat, 25 May 2013 08:11:18 -0500
> > > From: Stephen Politzer-Ahles < politzerahless at gmail.com >
> > > To: fieldtrip at donders.ru.nl
> > > Subject: [FieldTrip] Sourcespace and volume conductor misaligned
> > > Message-ID:
> > > <CAJT2k_9-hd_sM=hp4P-CUu+=aduSOMyZV7XPG0= fFk0ouR0wzA at mail.gmail.com
> > > >
> > > Content-Type: text/plain; charset="utf-8"
> > >
> > > Hello all,
> > >
> > > I am going through the workflow at
> > > http://fieldtrip.fcdonders.nl/tutorial/minimumnormestimate . After
> > > making
> > > the volume conduction model using ft_prepare_headmodel(), I noticed
> > > that
> > > although the volume conduction model and sourcespace have the same
> > > orientation and overall size/shape (after I converted the volume
> > > conduction
> > > model to cm, which wasn't in the tutorial but my original model came
> > > out in
> > > mm), they don't quite line up, as you can see in this figure:
> > >
> > > http://i.imgur.com/mGEtLOa.png
> > >
> > > I did interactively re-align the data to CTF (twice--in step 2 of
> > > "Preprocessing of the anatomical MRI" and in step 4 of "Source
> > > model")
> > > using fiducials, and to Talairach (step 5 of "Preprocessing of the
> > > anatomical data"), so I'm not sure how it ended up this way. The
> > > code I've
> > > used at each step is basically the same as that in the tutorial.
> > >
> > > Is there any way to line up my volume conduction model and
> > > sourcespace now,
> > > without going back and re-running most of the workflow?
> > >
> > > Best,
> > > Steve
> > >
> > > --
> > > Stephen Politzer-Ahles
> > > University of Kansas
> > > Linguistics Department
> > > http://people.ku.edu/~sjpa/
> > On Sat, May 25, 2013 at 1:56 PM, < fieldtrip-request at science.ru.nl >
> > wrote:
> > >
> > > Send fieldtrip mailing list submissions to
> > > fieldtrip at science.ru.nl
> > >
> > > To subscribe or unsubscribe via the World Wide Web, visit
> > > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
> > > or, via email, send a message with subject or body 'help' to
> > > fieldtrip-request at science.ru.nl
> > >
> > > You can reach the person managing the list at
> > > fieldtrip-owner at science.ru.nl
> > >
> > > When replying, please edit your Subject line so it is more specific
> > > than "Re: Contents of fieldtrip digest..."
> > >
> > >
> > > Today's Topics:
> > >
> > > 1. Sourcespace and volume conductor misaligned
> > > (Stephen Politzer-Ahles)
> > > 2. Re: fieldtrip Digest, Vol 30, Issue 31 (Johanna Zumer)
> > >
> > >
> > > ----------------------------------------------------------------------
> > >
> > > Message: 1
> > > Date: Sat, 25 May 2013 08:11:18 -0500
> > > From: Stephen Politzer-Ahles < politzerahless at gmail.com >
> > > To: fieldtrip at donders.ru.nl
> > > Subject: [FieldTrip] Sourcespace and volume conductor misaligned
> > > Message-ID:
> > > <CAJT2k_9-hd_sM=hp4P-CUu+=aduSOMyZV7XPG0= fFk0ouR0wzA at mail.gmail.com
> > > >
> > > Content-Type: text/plain; charset="utf-8"
> > >
> > > Hello all,
> > >
> > > I am going through the workflow at
> > > http://fieldtrip.fcdonders.nl/tutorial/minimumnormestimate . After
> > > making
> > > the volume conduction model using ft_prepare_headmodel(), I noticed
> > > that
> > > although the volume conduction model and sourcespace have the same
> > > orientation and overall size/shape (after I converted the volume
> > > conduction
> > > model to cm, which wasn't in the tutorial but my original model came
> > > out in
> > > mm), they don't quite line up, as you can see in this figure:
> > >
> > > http://i.imgur.com/mGEtLOa.png
> > >
> > > I did interactively re-align the data to CTF (twice--in step 2 of
> > > "Preprocessing of the anatomical MRI" and in step 4 of "Source
> > > model")
> > > using fiducials, and to Talairach (step 5 of "Preprocessing of the
> > > anatomical data"), so I'm not sure how it ended up this way. The
> > > code I've
> > > used at each step is basically the same as that in the tutorial.
> > >
> > > Is there any way to line up my volume conduction model and
> > > sourcespace now,
> > > without going back and re-running most of the workflow?
> > >
> > > Best,
> > > Steve
> > >
> > > --
> > > Stephen Politzer-Ahles
> > > University of Kansas
> > > Linguistics Department
> > > http://people.ku.edu/~sjpa/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20130601/1c615cda/attachment-0001.html>